Message Technologies Launches MTI Natural Chat

Message Technologies (MTI), yesterday released MTI Natural Chat, a hosted, on-demand service that allows enterprises to add intelligent voice search and natural language understanding (NLU) to customer service offerings.

In developing the solution, MTI partnered with several “best-of-breed” speech, Web, and artificial intelligence (AI) technology companies. Natural Chat allows any new or existing interactive voice response (IVR) application to incorporate an intelligent natural language option anywhere within the call flow.

With Natural Chat, a caller can speak a sentence describing his reason for calling. The spoken audio is streamed in real-time to a speech recognition engine, converted to a text string, and then applied to an interpretive engine that analyzes the text string using AI algorithms.

Once analyzed, the result is delivered to the caller using pre-recorded prompts or natural language text-to-speech (TTS) capabilities, eliminating the need to navigate complex menu structures.

“So you have a much friendlier front end,” says Darrell Knight, president of MTI. “And an accelerated back end.”

According to Knight, what makes Natural Chat different from other NLU technology is the systems ability to interpret converted utterances via AI.

“The [first level is] interpretation of the utterance and then the second level is interpretation of the results of the utterance,” he says. “And that second level is derived from the AI engine which is basing itself typically on application- or enterprise-specific data.”

MTI says Natural Chat will lead to significantly increased call completions, reduced per-call costs, and higher customer satisfaction and retention levels.

“People become more comfortable using [the IVR] because it’s more conversational and it’s giving them the answer more quickly," notes Knight, who says he anticipates the system reducing time spent in IVRs by as much as 30 percent to 50 percent.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues